Notes.ini Entry


Name:
    DAOS_Prune_Delete_Limit

Syntax
    DAOS_Prune_Delete_Limit=Number

Applies to:
    Servers

Add-on:

    First Release:
      8.5

    Obsolete since:

      Category:
        DAOS

      Default:
        No limit

      UI equivalent:
        None

      Description:
      This Notes.ini variable sets the maximum number of entries to delete in a single invocation of prune


      Running the Prune Command Manually - The daosmgr process runs prune at 2am nightly. The prune command can also be invoked manually through daosmgr. If the prune command is issued with no arguments, the response is the current setting of the deferred deletion interval. A numeric argument to prune is used as a temporary override to the deferred deletion interval value specified in the server doc. If a 'prune 0' is performed, all NLO files with a reference count of zero will be deleted. Deleting everything will recover as much space as possible, but it is somewhat risky, as it means that there will be NLO files that have not been backed up, and will therefore be unrecoverable.


      Diagnosing DAOS prune:

      To diagnose concerns with DAOS prune, consider the following possibilities:

      1) Prune Not Running - It's possible that prune is not running for some reason. Check the server log at the 2am automatic run time, and look for status messages from prune. If the catalog is not in the correct state, prune will not run. If there is an error message displayed, that may be a clue to the root cause.

      2) Run Prune Manually - If the 2am run does not have any useful output, prune can be invoked manually. The console command 'tell daosmgr prune' will display the current setting for the deferred deletion interval. Use that value to invoke prune again, and a prune operation will be performed: 'tell daosmgr prune 30' The prune operation will then display a message indicating what the outcome of the prune was. Note that a prune operation may take a significant amount of time if there are a large number of entries on the deletion list.

      3) Resync May Be Needed - If the prune operation indicates that it is not running because of the DAOS catalog state, a resync is necessary to correct that. Resync operations can take a significant amount of time, This should be scheduled for off-hours time that will not impact production operations. Once the catalog state is SYNCHRONIZED again, retry the prune operation.

      4) Fixup May Be Needed - If the resync operation is not completing, fixup may need to be run on NSF files that resync was unable to process properly. Correct any errors using fixup, and try the resync again.

      For related information about DAOS prune, see DAOS prune posts available from the Lotus Support Portal.

      How DAOS prune scans:

      There can be a significant number of entries on the deletion list, and traversing that list can take time. Because of this, DAOS prune releases the lock periodically so that other work does not get backlogged. When this is done, DAOS prune must start over at the beginning of the list to ensure that it visits all entries. This can cause the "number of entries analyzed" value to be significantly higher than the total number of entries on the list.

      Question
      After upgrading to a new fixpack/hotfix, why are "The DAOS catalog cannot be resynchronized because notes.ini enables HOTFIX" messages displayed?
      Answer

      An existing string constant was re-used in order to support some code changes as a hotfix and in fixpacks, so the resulting messages are a somewhat confusing. The messages added in this manner will display as:
      The DAOS catalog cannot be resynchronized because notes.ini enables HOTFIX - DAOS catalog state = Needs Resync. Pruning has been suspended until the DAOS Catalog is resynchronized.
      If the message includes the keyword HOTFIX, the prefix of the message up to that keyword should be ignored. In these cases, the message is related only to the DAOS prune operation, and the notes.ini settings are not actually causing a 'catalog cannot be resynchronized' problem.
      Caution: If the message does not include the HOTFIX keyword, and instead mentions only a notes.ini variable name, that is a legitimate problem that needs to be corrected.
      Messages that may be displayed in this manner:
      HOTFIX - DAOS catalog state = Needs Resync. Pruning has been suspended until the DAOS Catalog is resynchronized
      HOTFIX - Pruning has been suspended because one or more NSF files are not in either the SYNCHRONIZED or DELETED state. See "daosmgr dbsummary" output for details
      HOTFIX - DAOS prune deleted 33 objects and completed with error: (0) No error

      The corresponding fix in subsequent Domino releases uses a new string constant value, and does not have this problem. This behavior is strictly related to hotfix/fixpack versions of this fix.

      Note about the following message:
      HOTFIX - DAOS prune deleted 0 objects and completed with error: (0) No error
      This message indicates that the prune operation did not delete anything, but also did not encounter any errors. In most cases this is acceptable behavior. Typically some objects will be deleted during the daily prune run. If this 'deleted 0 objects' message is displayed daily for longer than the DAOS deferred deletion interval, this could indicate there is a problem with the prune process, and should be investigated.

      For more information on DAOS prune, see technote 1448358, "Understanding the DAOS prune operation "